Yor\`ub\'a is a widely spoken West African language with a writing system rich in tonal and orthographic diacritics. With very few exceptions, diacritics are omitted from electronic texts, due to limited device and application support. Diacritics provide morphological information, are crucial for lexical disambiguation, pronunciation and are vital for any Yor\`ub\'a text-to-speech (TTS), automatic speech recognition (ASR) and natural language processing (NLP) tasks. Reframing Automatic Diacritic Restoration (ADR) as a machine translation task, we experiment with two different attentive Sequence-to-Sequence neural models to process undiacritized text. On our evaluation dataset, this approach produces diacritization error rates of less than 5%. We have released pre-trained models, datasets and source-code as an open-source project to advance efforts on Yor\`ub\'a language technology.
Neural machine translation (NMT) has been a new paradigm in machine translation, and the attention mechanism has become the dominant approach with the state-of-the-art records in many language pairs. While there are variants of the attention mechanism, all of them use only temporal attention where one scalar value is assigned to one context vector corresponding to a source word. In this paper, we propose a fine-grained (or 2D) attention mechanism where each dimension of a context vector will receive a separate attention score. In experiments with the task of En-De and En-Fi translation, the fine-grained attention method improves the translation quality in terms of BLEU score. In addition, our alignment analysis reveals how the fine-grained attention mechanism exploits the internal structure of context vectors.
We present Marian, an efficient and self-contained Neural Machine Translation framework with an integrated automatic differentiation engine based on dynamic computation graphs. Marian is written entirely in C++. We describe the design of the encoder-decoder framework and demonstrate that a research-friendly toolkit can achieve high training and translation speed.
Although there are increasing and significant ties between China and Portuguese-speaking countries, there is not much parallel corpora in the Chinese-Portuguese language pair. Both languages are very populous, with 1.2 billion native Chinese speakers and 279 million native Portuguese speakers, the language pair, however, could be considered as low-resource in terms of available parallel corpora. In this paper, we describe our methods to curate Chinese-Portuguese parallel corpora and evaluate their quality. We extracted bilingual data from Macao government websites and proposed a hierarchical strategy to build a large parallel corpus. Experiments are conducted on existing and our corpora using both Phrased-Based Machine Translation (PBMT) and the state-of-the-art Neural Machine Translation (NMT) models. The results of this work can be used as a benchmark for future Chinese-Portuguese MT systems. The approach we used in this paper also shows a good example on how to boost performance of MT systems for low-resource language pairs.
Statistical machine translation (SMT) systems perform poorly when it is applied to new target domains. Our goal is to explore domain adaptation approaches and techniques for improving the translation quality of domain-specific SMT systems. However, translating texts from a specific domain (e.g., medicine) is full of challenges. The first challenge is ambiguity. Words or phrases contain different meanings in different contexts. The second one is language style due to the fact that texts from different genres are always presented in different syntax, length and structural organization. The third one is the out-of-vocabulary words (OOVs) problem. In-domain training data are often scarce with low terminology coverage. In this thesis, we explore the state-of-the-art domain adaptation approaches and propose effective solutions to address those problems.
One of the difficulties of neural machine translation (NMT) is the recall and appropriate translation of low-frequency words or phrases. In this paper, we propose a simple, fast, and effective method for recalling previously seen translation examples and incorporating them into the NMT decoding process. Specifically, for an input sentence, we use a search engine to retrieve sentence pairs whose source sides are similar with the input sentence, and then collect $n$-grams that are both in the retrieved target sentences and aligned with words that match in the source sentences, which we call "translation pieces". We compute pseudo-probabilities for each retrieved sentence based on similarities between the input sentence and the retrieved source sentences, and use these to weight the retrieved translation pieces. Finally, an existing NMT model is used to translate the input sentence, with an additional bonus given to outputs that contain the collected translation pieces. We show our method improves NMT translation results up to 6 BLEU points on three narrow domain translation tasks where repetitiveness of the target sentences is particularly salient. It also causes little increase in the translation time, and compares favorably to another alternative retrieval-based method with respect to accuracy, speed, and simplicity of implementation.
Earth observation (EO) sensors deliver data with daily or weekly temporal resolution. Most land use and land cover (LULC) approaches, however, expect cloud-free and mono-temporal observations. The increasing temporal capabilities of today's sensors enables the use of temporal, along with spectral and spatial features. Domains, such as speech recognition or neural machine translation, work with inherently temporal data and, today, achieve impressive results using sequential encoder-decoder structures. Inspired by these sequence-to-sequence models, we adapt an encoder structure with convolutional recurrent layers in order to approximate a phenological model for vegetation classes based on a temporal sequence of Sentinel 2 (S2) images. In our experiments, we visualize internal activations over a sequence of cloudy and non-cloudy images and find several recurrent cells, which reduce the input activity for cloudy observations. Hence, we assume that our network has learned cloud-filtering schemes solely from input data, which could alleviate the need for tedious cloud-filtering as a preprocessing step for many EO approaches. Moreover, using unfiltered temporal series of top-of-atmosphere (TOA) reflectance data, we achieved in our experiments state-of-the-art classification accuracies on a large number of crop classes with minimal preprocessing compared to other classification approaches.
This paper proposes an approach for applying GANs to NMT. We build a conditional sequence generative adversarial net which comprises of two adversarial sub models, a generator and a discriminator. The generator aims to generate sentences which are hard to be discriminated from human-translated sentences (i.e., the golden target sentences), And the discriminator makes efforts to discriminate the machine-generated sentences from human-translated ones. The two sub models play a mini-max game and achieve the win-win situation when they reach a Nash Equilibrium. Additionally, the static sentence-level BLEU is utilized as the reinforced objective for the generator, which biases the generation towards high BLEU points. During training, both the dynamic discriminator and the static BLEU objective are employed to evaluate the generated sentences and feedback the evaluations to guide the learning of the generator. Experimental results show that the proposed model consistently outperforms the traditional RNNSearch and the newly emerged state-of-the-art Transformer on English-German and Chinese-English translation tasks.
Neural Network has been successfully applied to many real-world problems, such as image recognition and machine translation. However, for the current architecture of neural networks, it is hard to perform complex cognitive tasks, for example, to process the image and audio inputs together. Cortex, as an important architecture in the brain, is important for animals to perform the complex cognitive task. We view the architecture of Cortex in the brain as a missing part in the design of the current artificial neural network. In this paper, we purpose Cortex Neural Network (CrtxNN). The Cortex Neural Network is an upper architecture of neural networks which motivated from cerebral cortex in the brain to handle different tasks in the same learning system. It is able to identify different tasks and solve them with different methods. In our implementation, the Cortex Neural Network is able to process different cognitive tasks and perform reflection to get a higher accuracy. We provide a series of experiments to examine the capability of the cortex architecture on traditional neural networks. Our experiments proved its ability on the Cortex Neural Network can reach accuracy by 98.32% on MNIST and 62% on CIFAR10 at the same time, which can promisingly reduce the loss by 40%.
In several recently proposed stochastic optimization methods (e.g. RMSProp, Adam, Adadelta), parameter updates are scaled by the inverse square roots of exponential moving averages of squared past gradients. Maintaining these per-parameter second-moment estimators requires memory equal to the number of parameters. For the case of neural network weight matrices, we propose maintaining only the per-row and per-column sums of these moving averages, and estimating the per-parameter second moments based on these sums. We demonstrate empirically that this method produces similar results to the baseline. Secondly, we show that adaptive methods can produce larger-than-desired updates when the decay rate of the second moment accumulator is too slow. We propose update clipping and a gradually increasing decay rate scheme as remedies. Combining these methods and dropping momentum, we achieve comparable results to the published Adam regime in training the Transformer model on the WMT 2014 English-German machine translation task, while using very little auxiliary storage in the optimizer. Finally, we propose scaling the parameter updates based on the scale of the parameters themselves.
Neural Machine Translation with its significant results, still has a great problem: lack or absence of parallel corpus for many languages. This article suggests a method for generating considerable amount of parallel corpus for any language pairs, extracted from open source materials existing on the Internet. Parallel corpus contents will be derived from video subtitles. It needs a set of video titles, with some attributes like release date, rating, duration and etc. Process of finding and downloading subtitle pairs for desired language pairs is automated by using a crawler. Finally sentence pairs will be extracted from synchronous dialogues in subtitles. The main problem of this method is unsynchronized subtitle pairs. Therefore subtitles will be verified before downloading. If two subtitle were not synchronized, then another subtitle of that video will be processed till it finds the matching subtitle. Using this approach gives ability to make context based parallel corpus through filtering videos by genre. Context based corpus can be used in complex translators which decode sentences by different networks after determining contents subject. Languages have many differences in their formal and informal styles, including words and syntax. Other advantage of this method is to make corpus of informal style of languages. Because most of movies dialogues are parts of a conversation. So they had informal style. This feature of generated corpus can be used in real-time translators to have more accurate conversation translations.
While most machine translation systems to date are trained on large parallel corpora, humans learn language in a different way: by being grounded in an environment and interacting with other humans. In this work, we propose a communication game where two agents, native speakers of their own respective languages, jointly learn to solve a visual referential task. We find that the ability to understand and translate a foreign language emerges as a means to achieve shared goals. The emergent translation is interactive and multimodal, and crucially does not require parallel corpora, but only monolingual, independent text and corresponding images. Our proposed translation model achieves this by grounding the source and target languages into a shared visual modality, and outperforms several baselines on both word-level and sentence-level translation tasks. Furthermore, we show that agents in a multilingual community learn to translate better and faster than in a bilingual communication setting.
Many Natural Language Processing and Computational Linguistics applications involves the generation of new texts based on some existing texts, such as summarization, text simplification and machine translation. However, there has been a serious problem haunting these applications for decades, that is, how to automatically and accurately assess quality of these applications. In this paper, we will present some preliminary results on one especially useful and challenging problem in NLP system evaluation: how to pinpoint content differences of two text passages (especially for large pas-sages such as articles and books). Our idea is intuitive and very different from existing approaches. We treat one text passage as a small knowledge base, and ask it a large number of questions to exhaustively identify all content points in it. By comparing the correctly answered questions from two text passages, we will be able to compare their content precisely. The experiment using 2007 DUC summarization corpus clearly shows promising results.
Relying entirely on an attention mechanism, the Transformer introduced by Vaswani et al. (2017) achieves state-of-the-art results for machine translation. In contrast to recurrent and convolutional neural networks, it does not explicitly model relative or absolute position information in its structure. Instead, it requires adding representations of absolute positions to its inputs. In this work we present an alternative approach, extending the self-attention mechanism to efficiently consider representations of the relative positions, or distances between sequence elements. On the WMT 2014 English-to-German and English-to-French translation tasks, this approach yields improvements of 1.3 BLEU and 0.3 BLEU over absolute position representations, respectively. Notably, we observe that combining relative and absolute position representations yields no further improvement in translation quality. We describe an efficient implementation of our method and cast it as an instance of relation-aware self-attention mechanisms that can generalize to arbitrary graph-labeled inputs.
The demo proposal presents a Phrase-based Sanskrit-Hindi (SaHiT) Statistical Machine Translation system. The system has been developed on Moses. 43k sentences of Sanskrit-Hindi parallel corpus and 56k sentences of a monolingual corpus in the target language (Hindi) have been used. This system gives 57 BLEU score.
Neural machine translation has achieved levels of fluency and adequacy that would have been surprising a short time ago. Output quality is extremely relevant for industry purposes, however it is equally important to produce results in the shortest time possible, mainly for latency-sensitive applications and to control cloud hosting costs. In this paper we show the effectiveness of translating with 8-bit quantization for models that have been trained using 32-bit floating point values. Results show that 8-bit translation makes a non-negligible impact in terms of speed with no degradation in accuracy and adequacy.
The recent popularity of deep neural networks (DNNs) has generated a lot of research interest in performing DNN-related computation efficiently. However, the primary focus is usually very narrow and limited to (i) inference -- i.e. how to efficiently execute already trained models and (ii) image classification networks as the primary benchmark for evaluation. Our primary goal in this work is to break this myopic view by (i) proposing a new benchmark for DNN training, called TBD (TBD is short for Training Benchmark for DNNs), that uses a representative set of DNN models that cover a wide range of machine learning applications: image classification, machine translation, speech recognition, object detection, adversarial networks, reinforcement learning, and (ii) by performing an extensive performance analysis of training these different applications on three major deep learning frameworks (TensorFlow, MXNet, CNTK) across different hardware configurations (single-GPU, multi-GPU, and multi-machine). TBD currently covers six major application domains and eight different state-of-the-art models. We present a new toolchain for performance analysis for these models that combines the targeted usage of existing performance analysis tools, careful selection of new and existing metrics and methodologies to analyze the results, and utilization of domain specific characteristics of DNN training. We also build a new set of tools for memory profiling in all three major frameworks; much needed tools that can finally shed some light on precisely how much memory is consumed by different data structures (weights, activations, gradients, workspace) in DNN training. By using our tools and methodologies, we make several important observations and recommendations on where the future research and optimization of DNN training should be focused.
Machine translation has recently achieved impressive performance thanks to recent advances in deep learning and the availability of large-scale parallel corpora. There have been numerous attempts to extend these successes to low-resource language pairs, yet requiring tens of thousands of parallel sentences. In this work, we take this research direction to the extreme and investigate whether it is possible to learn to translate even without any parallel data. We propose a model that takes sentences from monolingual corpora in two different languages and maps them into the same latent space. By learning to reconstruct in both languages from this shared feature space, the model effectively learns to translate without using any labeled data. We demonstrate our model on two widely used datasets and two language pairs, reporting BLEU scores of 32.8 and 15.1 on the Multi30k and WMT English-French datasets, without using even a single parallel sentence at training time.
Existing machine translation decoding algorithms generate translations in a strictly monotonic fashion and never revisit previous decisions. As a result, earlier mistakes cannot be corrected at a later stage. In this paper, we present a translation scheme that starts from an initial guess and then makes iterative improvements that may revisit previous decisions. We parameterize our model as a convolutional neural network that predicts discrete substitutions to an existing translation based on an attention mechanism over both the source sentence as well as the current translation output. By making less than one modification per sentence, we improve the output of a phrase-based translation system by up to 0.4 BLEU on WMT15 German-English translation.
Style transfer is the task of automatically transforming a piece of text in one particular style into another. A major barrier to progress in this field has been a lack of training and evaluation datasets, as well as benchmarks and automatic metrics. In this work, we create the largest corpus for a particular stylistic transfer (formality) and show that techniques from the machine translation community can serve as strong baselines for future work. We also discuss challenges of using automatic metrics.